block: add an explicit ->disk backpointer to the request_queue

Replace the magic lookup through the kobject tree with an explicit
backpointer, given that the device model links are set up and torn
down at times when I/O is still possible, leading to potential
NULL or invalid pointer dereferences.

Fixes: edb0872f44 ("block: move the bdi from the request_queue to the gendisk")
Reported-by: syzbot <syzbot+aa0801b6b32dca9dda82@syzkaller.appspotmail.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Tested-by: Sven Schnelle <svens@linux.ibm.com>
Link: https://lore.kernel.org/r/20210816134624.GA24234@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This commit is contained in:
Christoph Hellwig 2021-08-16 15:46:24 +02:00 committed by Jens Axboe
parent 61a35cfc26
commit d152c682f0
9 changed files with 26 additions and 26 deletions

View File

@ -5269,7 +5269,7 @@ bfq_set_next_ioprio_data(struct bfq_queue *bfqq, struct bfq_io_cq *bic)
switch (ioprio_class) {
default:
pr_err("bdi %s: bfq: bad prio class %d\n",
bdi_dev_name(queue_to_disk(bfqq->bfqd->queue)->bdi),
bdi_dev_name(bfqq->bfqd->queue->disk->bdi),
ioprio_class);
fallthrough;
case IOPRIO_CLASS_NONE:

View File

@ -489,9 +489,9 @@ static int blkcg_reset_stats(struct cgroup_subsys_state *css,
const char *blkg_dev_name(struct blkcg_gq *blkg)
{
if (!queue_has_disk(blkg->q) || !queue_to_disk(blkg->q)->bdi->dev)
if (!blkg->q->disk || !blkg->q->disk->bdi->dev)
return NULL;
return bdi_dev_name(queue_to_disk(blkg->q)->bdi);
return bdi_dev_name(blkg->q->disk->bdi);
}
/**

View File

@ -525,7 +525,7 @@ void blk_mq_free_request(struct request *rq)
__blk_mq_dec_active_requests(hctx);
if (unlikely(laptop_mode && !blk_rq_is_passthrough(rq)))
laptop_io_completion(queue_to_disk(q)->bdi);
laptop_io_completion(q->disk->bdi);
rq_qos_done(q, rq);

View File

@ -141,9 +141,9 @@ void blk_queue_max_hw_sectors(struct request_queue *q, unsigned int max_hw_secto
limits->logical_block_size >> SECTOR_SHIFT);
limits->max_sectors = max_sectors;
if (!queue_has_disk(q))
if (!q->disk)
return;
queue_to_disk(q)->bdi->io_pages = max_sectors >> (PAGE_SHIFT - 9);
q->disk->bdi->io_pages = max_sectors >> (PAGE_SHIFT - 9);
}
EXPORT_SYMBOL(blk_queue_max_hw_sectors);
@ -475,9 +475,9 @@ EXPORT_SYMBOL(blk_limits_io_opt);
void blk_queue_io_opt(struct request_queue *q, unsigned int opt)
{
blk_limits_io_opt(&q->limits, opt);
if (!queue_has_disk(q))
if (!q->disk)
return;
queue_to_disk(q)->bdi->ra_pages =
q->disk->bdi->ra_pages =
max(queue_io_opt(q) * 2 / PAGE_SIZE, VM_READAHEAD_PAGES);
}
EXPORT_SYMBOL(blk_queue_io_opt);

View File

@ -90,9 +90,9 @@ static ssize_t queue_ra_show(struct request_queue *q, char *page)
{
unsigned long ra_kb;
if (!queue_has_disk(q))
if (!q->disk)
return -EINVAL;
ra_kb = queue_to_disk(q)->bdi->ra_pages << (PAGE_SHIFT - 10);
ra_kb = q->disk->bdi->ra_pages << (PAGE_SHIFT - 10);
return queue_var_show(ra_kb, page);
}
@ -102,12 +102,12 @@ queue_ra_store(struct request_queue *q, const char *page, size_t count)
unsigned long ra_kb;
ssize_t ret;
if (!queue_has_disk(q))
if (!q->disk)
return -EINVAL;
ret = queue_var_store(&ra_kb, page, count);
if (ret < 0)
return ret;
queue_to_disk(q)->bdi->ra_pages = ra_kb >> (PAGE_SHIFT - 10);
q->disk->bdi->ra_pages = ra_kb >> (PAGE_SHIFT - 10);
return ret;
}
@ -254,9 +254,8 @@ queue_max_sectors_store(struct request_queue *q, const char *page, size_t count)
spin_lock_irq(&q->queue_lock);
q->limits.max_sectors = max_sectors_kb << 1;
if (queue_has_disk(q))
queue_to_disk(q)->bdi->io_pages =
max_sectors_kb >> (PAGE_SHIFT - 10);
if (q->disk)
q->disk->bdi->io_pages = max_sectors_kb >> (PAGE_SHIFT - 10);
spin_unlock_irq(&q->queue_lock);
return ret;

View File

@ -97,7 +97,7 @@ static void wb_timestamp(struct rq_wb *rwb, unsigned long *var)
*/
static bool wb_recent_wait(struct rq_wb *rwb)
{
struct bdi_writeback *wb = &queue_to_disk(rwb->rqos.q)->bdi->wb;
struct bdi_writeback *wb = &rwb->rqos.q->disk->bdi->wb;
return time_before(jiffies, wb->dirty_sleep + HZ);
}
@ -234,7 +234,7 @@ enum {
static int latency_exceeded(struct rq_wb *rwb, struct blk_rq_stat *stat)
{
struct backing_dev_info *bdi = queue_to_disk(rwb->rqos.q)->bdi;
struct backing_dev_info *bdi = rwb->rqos.q->disk->bdi;
struct rq_depth *rqd = &rwb->rq_depth;
u64 thislat;
@ -287,7 +287,7 @@ static int latency_exceeded(struct rq_wb *rwb, struct blk_rq_stat *stat)
static void rwb_trace_step(struct rq_wb *rwb, const char *msg)
{
struct backing_dev_info *bdi = queue_to_disk(rwb->rqos.q)->bdi;
struct backing_dev_info *bdi = rwb->rqos.q->disk->bdi;
struct rq_depth *rqd = &rwb->rq_depth;
trace_wbt_step(bdi, msg, rqd->scale_step, rwb->cur_win_nsec,
@ -359,8 +359,8 @@ static void wb_timer_fn(struct blk_stat_callback *cb)
status = latency_exceeded(rwb, cb->stat);
trace_wbt_timer(queue_to_disk(rwb->rqos.q)->bdi, status,
rqd->scale_step, inflight);
trace_wbt_timer(rwb->rqos.q->disk->bdi, status, rqd->scale_step,
inflight);
/*
* If we exceeded the latency target, step down. If we did not,

View File

@ -1078,6 +1078,7 @@ static void disk_release(struct device *dev)
disk_release_events(disk);
kfree(disk->random);
xa_destroy(&disk->part_tbl);
disk->queue->disk = NULL;
blk_put_queue(disk->queue);
iput(disk->part0->bd_inode); /* frees the disk */
}
@ -1276,6 +1277,7 @@ struct gendisk *__alloc_disk_node(struct request_queue *q, int node_id,
device_initialize(disk_to_dev(disk));
inc_diskseq(disk);
disk->queue = q;
q->disk = disk;
lockdep_init_map(&disk->lockdep_map, "(bio completion)", lkclass, 0);
#ifdef CONFIG_BLOCK_HOLDER_DEPRECATED
INIT_LIST_HEAD(&disk->slave_bdevs);

View File

@ -421,6 +421,8 @@ struct request_queue {
spinlock_t queue_lock;
struct gendisk *disk;
/*
* queue kobject
*/
@ -661,9 +663,6 @@ extern void blk_clear_pm_only(struct request_queue *q);
dma_map_page_attrs(dev, (bv)->bv_page, (bv)->bv_offset, (bv)->bv_len, \
(dir), (attrs))
#define queue_has_disk(q) ((q)->kobj.parent != NULL)
#define queue_to_disk(q) (dev_to_disk(kobj_to_dev((q)->kobj.parent)))
static inline bool queue_is_mq(struct request_queue *q)
{
return q->mq_ops;

View File

@ -30,7 +30,7 @@ TRACE_EVENT(kyber_latency,
),
TP_fast_assign(
__entry->dev = disk_devt(queue_to_disk(q));
__entry->dev = disk_devt(q->disk);
strlcpy(__entry->domain, domain, sizeof(__entry->domain));
strlcpy(__entry->type, type, sizeof(__entry->type));
__entry->percentile = percentile;
@ -59,7 +59,7 @@ TRACE_EVENT(kyber_adjust,
),
TP_fast_assign(
__entry->dev = disk_devt(queue_to_disk(q));
__entry->dev = disk_devt(q->disk);
strlcpy(__entry->domain, domain, sizeof(__entry->domain));
__entry->depth = depth;
),
@ -81,7 +81,7 @@ TRACE_EVENT(kyber_throttled,
),
TP_fast_assign(
__entry->dev = disk_devt(queue_to_disk(q));
__entry->dev = disk_devt(q->disk);
strlcpy(__entry->domain, domain, sizeof(__entry->domain));
),